Clojurians Log v2

Everything feels fine when tests pass.

btw, I put foo bar above but a real result is:

expected: (= (testfn test) (expected test))
  actual: (not (= "{:type \"Mapping\",\n :coll\n ({:type \"Scalar\", :style \":\", :value \"a\"}\n  {:type \"Scalar\", :style \":\", :value \"b c\"})}\n" "{:type \"Mapping\",\n :coll\n ({:type \"Scalar\", :style \":\", :value \"a\"}\n  {:type \"Scalar\", :style \":\", :value \"b c\"})}X\n"))

When I find there's too much output I just use a let binding before the test expression

You might like https://github.com/pjstadig/humane-test-output which improves the reporting in a lot of cases.

@afoltzm not sure I follow. Example?

@seancorfield I'll take a look

I just don't get why it's putting the testing code in expected My testing function produces a string, so I want expected to be the string I'm declaring to be expected, and actual to be the output of my testing function.

The actual expected and actual values are printed in the second line

clojure.test is pretty simplistic. If you say (is some-expression) then that is the "expected" expression, and the "actual" result is (not value-of-some-expression).

The way it's printed is a bit unorthodox but you get used to it

lol

That's why things like HTO got written and why there are also other test runners, like Kaocha, and other test libraries, like Expectations.

HTO looks promising

I use https://github.com/pesterhazy/elucidate when a failure is hard to read

I liked Expectations enough that I took over maintenance of it 🙂 and I recommend HTO even with Expectations. But clojure.test's extension hooks are very primitive so there's a limit to what test runners can actually do.

I made a really nice data driven testing lib over clojure.test, so I need something that works with clojure.test

HTO does, right?

Expectations has two versions: the original standalone version and the clojure.test-compatible version. And, yes, HTO is an extension for clojure.test.

I like eftest nowadays. The defaults are quite nice, it’s simple and writing a custom runner or reporter is trivial

https://github.com/weavejester/eftest/blob/master/doc/failing-example.png

Not to mention it can run tests in parallel (per var or ns)

HTO is exactly what I wanted here! And trivial to add. No changes to my setup.

Testing yamlscript.composer-test
* test-1: Example Test

FAIL in (test-1) (test.clj:114)
expected: {:type "Mapping",
           :coll
           ({:type "Scalar", :style ":", :value "a"}
            {:type "Scalar", :style ":", :value "b c"})}
  actual: {:type "Mapping",
           :coll
           ({:type "Scalar", :style ":", :value "a"}
            {:type "Scalar", :style ":", :value "b C"})}
    diff: - {:coll [nil {:value "b c"}]}
          + {:coll [nil {:value "b C"}]}

Ran 1 tests containing 1 assertions.
1 failures, 0 errors.
{:test 1, :pass 0, :fail 1, :error 0, :type :summary}

Thanks all. I'll check out the others someday 🙂

https://github.com/lambdaisland/kaocha does nice reporting too. It is more of a complete replacement runner, so some folks find that too much, but I tend to like it!

And when you get into diffing more complex structures, I think that like me, you'll find https://github.com/nubank/matcher-combinators a complete joy.

I really, really recommend matcher-combinators, even for simple things

I quite like it as well